Body Information Acquisition System of Redundant Musculo-Skeletal Humanoid

Abstract

We propose a design of musculo-skeletal humanoid whose muscles can be easily added, removed, or re-arranged, which can be compared to the ability of growing up of a human’s (or an animal’s) body. This paper presents a design and development of the reconfigurable complex-body robot, and a method to build a system which can acquire the body information and can be used for realization of motions of human-like nonlinear musculo-skeletal robots. It shows that even reconfigurable complex-body robot’s motion can be achieved by a feedback modification method in conjunction with neural network. We did basic validation experiments by giving target postures one by one, measuring the realized postures, and comparing them with the target postures. We also carried out whole-body motion experiments using motion captured sequences using the presented system.